In [2]:
"moham_Task".replace("_"," ").upper()
Out[2]:
'MOHAM TASK'

Melbourne House Prices EDA Book: A guide on analysing housing data in Melbourne, Victoria.¶

Melbourne Housing

Insights about Melbourne city to helps us build a holistic view.¶

Melbourne is home to the second most expensive housing market in Australia after Sydney.

Even the 25th percentile has house sale prices at $665,000

Dallas and Broadmeadows are home to some of the cheapest housing in Melbourne

Reference; https://thepropertytribune.com.au/market-insights/what-are-the-cheapest-houses-in-melbourne-to-buy/

Step1 : Importing Libaries¶

In [1]:
#Library
import numpy as np 
import pandas as pd 
import seaborn as sns
import datetime
import time
import matplotlib.pyplot as plt
import plotly.graph_objs as go
import plotly.offline as py
from plotly.offline import download_plotlyjs, init_notebook_mode, iplot

# py.init_notebook_mode(connected = True)
pd.set_option('display.float_format', lambda x: f'{x:.1f}')
df = pd.read_csv("Melbourne_housing_2000data_WilsonWU.csv")

First 5 Rows of the data¶

In [2]:
df.head()
Out[2]:
Price Distance Bedroom Bathroom CarSpace Landsize BuildingArea YearBuilt Regionname Date Lattitude Longtitude
0 2550000 6.8 5 2 2 840 244.0 1910 Western Metropolitan 19/08/2017 -37.9 144.9
1 2325000 7.4 5 3 2 501 233.0 2006 Southern Metropolitan 25/02/2017 -37.9 145.0
2 1250000 1.6 3 1 1 113 105.0 1890 Northern Metropolitan 17/09/2016 -37.8 145.0
3 953000 15.0 4 2 2 551 150.0 1997 Northern Metropolitan 15/10/2016 -37.7 145.1
4 2200000 8.4 3 2 2 735 224.0 1920 Southern Metropolitan 24/02/2018 -37.9 145.1

Data Cleaning¶

In [3]:
#Categorical Data
df['Regionname'] = df['Regionname'].astype('category')



#Integer Data
df['Bedroom'] = df['Bedroom'].astype('int64')
df['Bathroom'] = df['Bathroom'].astype('int64')
df['CarSpace'] = df['CarSpace'].astype('int64')
df['Bathroom'] = df['Bathroom'].astype('int64')
df['Landsize'] = df['Landsize'].astype('int64')
df['YearBuilt'] = df['YearBuilt'].astype('int64')
df['BuildingArea'] = df['BuildingArea'].astype('int64')


df['Price'] = df['Price'].fillna(0)
df['Price'] = df['Price'].astype('float64')
df['Distance'] = df['Distance'].astype('float64')
df['Lattitude'] = df['Lattitude'].astype('float64')
df['Longtitude'] = df['Longtitude'].astype('float64')



#Date Data
df['Date'] = pd.to_datetime(df['Date'],format='%d/%m/%Y')
df.rename(columns = {'Date':'DateSold'}, inplace = True)
df.info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 2000 entries, 0 to 1999
Data columns (total 12 columns):
 #   Column        Non-Null Count  Dtype         
---  ------        --------------  -----         
 0   Price         2000 non-null   float64       
 1   Distance      2000 non-null   float64       
 2   Bedroom       2000 non-null   int64         
 3   Bathroom      2000 non-null   int64         
 4   CarSpace      2000 non-null   int64         
 5   Landsize      2000 non-null   int64         
 6   BuildingArea  2000 non-null   int64         
 7   YearBuilt     2000 non-null   int64         
 8   Regionname    2000 non-null   category      
 9   DateSold      2000 non-null   datetime64[ns]
 10  Lattitude     2000 non-null   float64       
 11  Longtitude    2000 non-null   float64       
dtypes: category(1), datetime64[ns](1), float64(4), int64(6)
memory usage: 174.3 KB

Price Heat Map Across Regions¶

In [4]:
import plotly.express as px
fig = px.density_mapbox(df, lat='Lattitude', lon='Longtitude', z='Price', radius=10,
                        center=dict(lat=-37.8, lon=145), zoom=10,
                        mapbox_style="stamen-terrain", opacity = 0.5, title = 'Melbourne Price Heatmap')
# fig.show()
Map tiles by Stamen Design under CC BY 3.0 | Data by OpenStreetMap contributors under ODbL
1M2M3M4M5M6M7M8MPriceMelbourne Price Heatmap
plotly-logomark

Solving Business Questions using EDA¶

  1. What region is the most expensive in Melbourn?
  2. Are older houses worth more or less?
  3. What factors effect Melbourne House prices most?

Q1. What region is the most expensive in Melbourne?¶

To answer this question, we can group the data by 'Regionname' and calculate the mean price for each region:

In [5]:
import plotly.express as px
df_region = df.groupby('Regionname').agg({'Price': 'mean'}).reset_index().sort_values(by = ['Price'], ascending=True) #price ->  'house selling price'
fig1 = px.bar(df_region, x='Price', y='Regionname', orientation='h', title='Mean Price by Region')
fig1.show()
00.2M0.4M0.6M0.8M1M1.2M1.4M1.6MWestern VictoriaNorthern VictoriaEastern VictoriaWestern MetropolitanSouth-Eastern MetropolitanNorthern MetropolitanEastern MetropolitanSouthern Metropolitan
Mean Price by RegionPriceRegionname
plotly-logomark

Inference from the above bar Plot ;¶

According to the plot

  • Southern Metropolitan have tha highest average selling price.
  • Western Victoria have tha highest average selling price.

Q2. Are older houses worth more or less?¶

To answer this question, First We create a new columne; Age which is Year of DateSold and YearBuilt and then we can create a scatterplot of 'Price' against 'Age'

In [6]:
fig2 = px.scatter(df, x='YearBuilt', y='Price', title='Price vs YearBuilt')
fig2.show()
# This code will create a scatter plot showing the relationship between 'Price' and 'YearBuilt':
12001300140015001600170018001900200001M2M3M4M5M6M7M8M
Price vs YearBuiltYearBuiltPrice
plotly-logomark

Inference from the above scatter Plot ;¶

According to the plot, there doesn't seem to be a strong relationship between the year a house was built and its price. However, there are some outliers, particularly for houses built before 1900, which seem to have higher prices.

So, Let's kick outlier and Check again.

In [7]:
df[['DateSold','YearBuilt']].info()

<class 'pandas.core.frame.DataFrame'>
RangeIndex: 2000 entries, 0 to 1999
Data columns (total 2 columns):
 #   Column     Non-Null Count  Dtype         
---  ------     --------------  -----         
 0   DateSold   2000 non-null   datetime64[ns]
 1   YearBuilt  2000 non-null   int64         
dtypes: datetime64[ns](1), int64(1)
memory usage: 31.4 KB

We can check dtype using info() for dataframe¶

Task : To calculate Age of House at the time of House Sold Date

We can check 'DateSold' is 'datetime64' so using dt.year we can have extract 'Year' and 'YearBuilt' is already int.

In [8]:
df['Age'] = df['DateSold'].dt.year - df['YearBuilt']
df.head()
Out[8]:
Price Distance Bedroom Bathroom CarSpace Landsize BuildingArea YearBuilt Regionname DateSold Lattitude Longtitude Age
0 2550000.0 6.8 5 2 2 840 244 1910 Western Metropolitan 2017-08-19 -37.9 144.9 107
1 2325000.0 7.4 5 3 2 501 233 2006 Southern Metropolitan 2017-02-25 -37.9 145.0 11
2 1250000.0 1.6 3 1 1 113 105 1890 Northern Metropolitan 2016-09-17 -37.8 145.0 126
3 953000.0 15.0 4 2 2 551 150 1997 Northern Metropolitan 2016-10-15 -37.7 145.1 19
4 2200000.0 8.4 3 2 2 735 224 1920 Southern Metropolitan 2018-02-24 -37.9 145.1 98

Removing Outlier => Removing data where YearBuilt Less than 1900 ;¶

In [9]:
df_cleaned = df[df['YearBuilt'] >= 1900]
fig2 = px.scatter(df_cleaned, x='YearBuilt', y='Price', title='Price vs YearBuilt')
fig2.show()
# This code will create a scatter plot showing the relationship between 'Price' and 'YearBuilt':
190019201940196019802000202001M2M3M4M5M6M
Price vs YearBuiltYearBuiltPrice
plotly-logomark
In [10]:
# Q2: Are older houses worth more or less?
df_cleaned = df[df['YearBuilt'] > 1800]
fig2 = px.scatter(df_cleaned, x='Age', y='Price', title='Price vs Age')
fig2.show()
02040608010012014016001M2M3M4M5M6M7M8M
Price vs AgeAgePrice
plotly-logomark

Q3. What factors affect Melbourne house prices most?¶

To answer this question, To answer this question, we can create a correlation matrix using the px.imshow() function in Plotly:

In [11]:
# Using Seaborn Library
# Create heatmap

corr_matrix = df.corr()
fig, ax = plt.subplots(figsize=(10,8))
sns.heatmap(corr_matrix, annot=True, cmap='coolwarm', ax=ax)

# Customize plot
ax.set_title("Correlation Matrix for Melbourne Housing Data", fontsize=16)
ax.tick_params(axis='x', labelsize=12, rotation=45)
ax.tick_params(axis='y', labelsize=12, rotation=0)

# Show the plot
plt.show()

Inference from the Corr Chart Price with other Cols¶

According to the correlation matrix, the variables that have the highest correlation with 'Price' are 'Rooms' (0.5), 'Bathroom' (0.46), 'BuildingArea' (0.53), and 'Distance' (-0.21).

This suggests that the size of the property (measured by the number of rooms, bathrooms, and building area) and its location (measured by the distance to the city center) are the most important factors affecting house prices in Melbourne.